Gibbsite crystal agglomeration is a critical size and morphology control operation in the commercially important Bayer process for producing alumina. There has long been a need for a mathematical gibbsite agglomeration model that is both capable of describing the process and is easy to implement. In this paper, a simple model is presented for the agglomeration efficiency of gibbsite crystals during precipitation from seeded sodium aluminate solutions in a turbulently stirred vessel. The model has been developed by viewing agglomeration as a combination of reversible and irreversible pseudo-reaction steps, involving the sub-processes of particle transport, collisions, capture, aggregate rupture, and aggregate cementation. A range of models for these sub-processes were trialled. The model was tuned using gibbsite precipitation data from an extensive experimental program covering a wide range of temperatures, supersaturations, particle sizes, shear rates and solids concentrations relevant to the Bayer process. The resultant agglomeration efficiency model is a function of liquor viscosity, gibbsite crystal growth rate, particles size, impeller tip speed and fluid shear rate. The effects of the key process variables of liquor composition, operating temperature and hydrodynamics are incorporated via these terms. The agglomeration efficiency model was validated by incorporating it into a population balance model of a laboratory precipitator and demonstrating good agreement between the simulated dynamic particle size distributions and the experimental values. Crown Copyright (c) 2005 Published by Elsevier Ltd. All rights reserved.